Why can there be several seconds of lag between a binary neutron star merger and the emission of gamma rays from the same area?

Question

Answers to: Gravitational waves and gamma ray burst: how were the error bars determined for this speed of gravity calculation? Was $H_0$ used? go into substantial detail explaining this determination. These have been helpful, which encouraged me to try to read the linked paper again. However, based on comments and the paper itself, I see that assigning limits to what sounds to me to be an emission lag of several seconds between the end of the merger and a subsequent gamma ray emission from the same general object/area had to be at least considered and included in some way.

Here I'd like to ask only about this lag. While I'd also like to know what produces the time-lagged gamma ray emission I'm not sure a priori if that is a different question and/or if it's been answered already.

Question: Why can there be several seconds of lag between a binary neutron star merger and the emission of gamma rays from the same region?

Answer 1

The paper (section 5.1) discusses three possibilities in the context of a relativistic fireball model, where some of the kinetic energy in relativistic jets of material emerging from the explosion is converted into gamma rays. The gravitational wave emission is always "prompt" since any surrounding material is transparent to gravitational waves. In contrast, the gamma ray emission may be delayed because:

1. The gamma rays are produced by shocks. There is a delay in the jets reaching the radius where the shocks form, either generated internally in the fireball by components with different speeds or just through deceleration in the interstellar medium.

2. The fireball is initially opaque to gamma rays and it is not until it has expanded sufficiently that it becomes transparent, allowing the gamma rays to escape.

3. It may take some time for the jet to punch out of the dense inner regions of the explosion.